Operational Amplifier Noise Prediction

(All Op Amps)
Application Note

Introduction
When working with op amp circuits an engineer is frequently
required to predict the total RMS output noise in a given
bandwidth for a certain feedback configuration. While op
amp noise can be expressed in a number of ways, spot
noise (RMS input voltage noise or current noise which
would pass through 1Hz wide bandpass filters centered at
various discrete frequencies), affords a universal method of
predicting output noise in any op amp configuration.

The Noise Model

November 1996

AN519.1

Procedure for Computing Total Output

Noise
1. Refer to the voltage noise curves for the amplifier to be
used.
2. Enter values of eni2 line (a) of the table below from the
curve labeled Noise spectral density (the values must
be squared).
3. From the current noise curves for the amplifier, obtain the
values of ini2 for each of the frequencies in the table,
and multiply each by RG2, entering the products in line
(b) of the table.

Figure 1 is a typical noise model depicting the noise voltage

and noise current sources that are added together in the
form of root mean square to give the total equivalent input
voltage noise (RMS), therefore:

4. Obtain the value of 4KTRG from Figure 8, and enter it on

line (c) of the table. This is constant for all frequencies.
The 4KTRG value must be adjusted for temperatures
other than normal room temperature.

E ni =

5. Total each column in the table on line (d). This total is

Eni2.

e 2 + I 2 R 2 + 4KTR G where,
ni
ni G

Eni is the total equivalent input voltage noise of the circuit,

10Hz

eni is the equivalent input voltage noise of the amplifier, and

(a) eni2

lni2RG2 is the voltage noise generated by the current noise.

(b)
Ini2RG2

4KTRG expresses the thermal noise generated by the external

resistors in the circuit where K = 1.38 x 10-23 joules/oK;
T = 300oK (27oC) and:
R1 R3
R G = --------------------- + R 2
R 1 + R 3
4KTR1

R3
A
+

eni *
4KTR2
ini1

1kHz

10kHz

100kHz

(c) 4KTRG
(d) Eni2

6. On linear scale graph paper enter each of the values for

Eni2 versus frequency. In most cases, sufficient accuracy
can be obtained simply by joining the points on the graph
with straight line segments.

4KTR3

R1

100Hz

ini2

R2

FIGURE 1.

The total RMS output noise (Eno) of an amplifier stage with

gain = G in the bandwidth between f1 and f2 is:
f2
2 1 2
E no = G ( E ni ) df

f1

Note that in the amplifier stage shown, G is the non-inverting

gain:
R 3

G = 1 + ------- ,
R 1

For the bandwidth of interest, calculate the area under the

curve by adding the areas of trapezoidal segments. This
procedure assumes a perfectly square bandpass condition;
to allow for the more normal -6dB/octave bandpass skirts,
multiply the upper (-3dB) frequency by 1.57 to obtain the
effective bandwidth of the circuit, before computing the area.
The total area obtained is equivalent to the square of the
total input noise over the given bandwidth.
7. Take the square root of the area found above and multiply
by the gain (G) of the circuit to find the total Output RMS
noise.
8. Take the square root of the area found above and multiply
by the gain (G) of the circuit to find the total Output RMS
noise.

A Typical Example
It is necessary to find the output noise of the circuit shown
below between 1kHz and 24kHz.

regardless of which input is normally driven.

1-888-INTERSIL or 321-724-7143 | Copyright

Intersil Corporation 1999

Application Note 519

Actual Measurements for Comparison

15k
+

The circuit shown below was used to actually measure the

broadband noise of the HA-2600 for the selected bandwidth:

HA-2600

15k

100

0.1F

15k

100

100

15k

FIGURE 2. THE HA-2600 IN A TYPICAL G = 1000 CIRCUIT

Values are selected from the data sheet and Figure 8 to fill in
the table as shown below. An RG of 30k was selected.
10Hz

100Hz

1kHz

10kHz

100kHz

(a) eni2

3.6x10-15

1.156x10-15 7.84x10-16

7.29x10-16

7.29x10-16

(b)
Ini2RG2

9.9x10-16

1.89x10-16

7.2x10-18

72x10-18

3.15x10-17

(c) 4KTRG 4.968x10-16 4.968x10-16 4.968x10-16 4.968x10-16 4.968x10-16

(d) Eni2

5.09 x10-15

1.86x10-15

1.31x10-15

1.23x10-15

1.23x10-15

The totals of the selected values for each frequency is in the

form of Eni2. This should be plotted on linear graph paper as
shown below:
1.5 x 10-15

1.5k

100

FIGURE 4. A TYPICAL TEST CIRCUIT FOR BROADBAND

NOISE MEASUREMENTS

The frequencies below the f1 point of the bandwidth selected

are filtered out by the RC network on the output of HA-2600.
The measurement of the broadband noise is observed on
the true RMS voltmeter. The measured output noise of the
circuit is 4.7mVRMS as compared to the calculated value of
6.7mVRMS.

Acquiring the Data for Calculations

Spot noise values must be generated in order to make the
output noise prediction. The effects of Popcorn noise have
been excluded due to the type of measurement system.
The Quan-Tech Control Unit, Model No. 2283 and Filter Unit,
Model No. 2181 were used to acquire spot noise voltage
values expressed in (V/Hz). The test system performs
measurements from 10Hz by orders of magnitude to 100kHz
with an effective bandwidth of 1Hz at each tested frequency.

(f1)
f2 (1.57)

1.0 x 10-15
Ein2, (V2/Hz)

TRUE
RMS
METER

HA-2600

Several source resistance (RG) values were used in the

measuring system to reveal the effects of RG on each type
of Intersil's op amps and to obtain proper voltage noise values essential for current noise calculations.

0.5 x 10-15

A Discussion On Popcorn Noise

1K

10K

20K
30K
FREQUENCY (Hz)

40K

50K

FIGURE 3. HA-2600 TOTAL EQUIVALENT INPUT NOISE

Since a noise figure is needed for the frequency of 1kHz to

24kHz, it is necessary to calculate the effective bandwidth of
the circuit. With AV = 60dB the upper 3dB point is approximately 24kHz. The product of 1.57 (24kHz) is 37.7kHz and
is the effective bandwidth of the circuit.
The shaded area under the curve is approximately
45 x 10-12 V2; the total equivalent input noise is Ein2 or
6.7V, and the total output noise for the selected bandwidth
is Ein2 x (closed loop gain) or 6.7mVRMS.

Popcorn noise was first discovered in early 709 type op

amps. Essentially it is an abrupt step-like shift in offset
voltage (or current) lasting for several milliseconds and
having amplitude from less than one microvolt to several
hundred microvolts. Occurrence of the pops is quite
random - an amplifier may exhibit several pops per second
during one observation period and then remain popless for
several minutes. Worst case conditions are usually at low
temperatures with high values of RG. Some amplifier
designs and some manufacturer's products are notoriously
bad in this respect. Although theories of the popcorn mechanism differ, it is known that devices with surface contamination of the semiconductor chip will be particularly bad
poppers. Advertising claims not withstanding, the author
has never seen any manufacturer's op amp that was completely free of popcorn. Some peak detector circuits have
been developed to screen devices for low amplitude pops,
but 100% assurance is impossible because an infinite test
time would be required. Some studies have shown that spot
noise measurements at 10Hz and 100Hz, discarding units
that are much higher than typical, is an effective screen for
potentially high popcorn units.

Application Note 519

References

The vast majority of Intersil op amps will exhibit less than

3VP-P popcorn. Screening can be performed, but it
should be noted that the confidence level of the screen could
be as low as 60%.

[1] Fitchen, F.C. and Motchenbacker, C.D. Low Noise

Electronic Design. New York: John Wiley and Sons,
1973.
Instruction Manual, Model 2173C Transistor Noise Analyzer
Control Unit. Quan-Tech, Division of KMS Industries. Whippany, New Jersey.

Typical Spot Noise Curves

Unless Otherwise Noted: VS = 15V, TA = 25oC

MEAN SQUARE NOISE CURRENT (A2/Hz)

SOURCE RESISTANCE = 300k

SOURCE RESISTANCE = 100k

10-12
10-13
10-14
10-15

10-16
10

SOURCE RESISTANCE = 0

100

1K
10K
FREQUENCY (Hz)

10-22

10-23

10-24

10-25

10-26
10

100K

100

1K
FREQUENCY (Hz)

10K

100K

FIGURE 6. HA-2500/2510/2520 INPUT NOISE CURRENT

10-14

10-24

MEAN SQUARE NOISE CURRENT (A2/Hz)

FIGURE 5. HA-2500/2510/2520 INPUT NOISE VOLTAGE

SOURCE RESISTANCE = 100k

10-15

SOURCE RESISTANCE = 0
10-16

10-17
10

100

1K
FREQUENCY (Hz)

10K

100K

10-25

10-26

10-27
10

FIGURE 7. HA-4741 INPUT NOISE VOLTAGE

100

1K
FREQUENCY (Hz)

10-13
10-14
4KTRG
10-15

10-16

10-17
1K

10K

100K

1M

10M

RG ()

FIGURE 9. NOISE vs RESISTOR VALUE

10K

FIGURE 8. HA-4741 INPUT NOISE CURRENT

10-12

4KTRG

MEAN SQUARE NOISE VOLTAGE (V2/Hz)

MEAN SQUARE NOISE VOLTAGE (V2/Hz)

10-21
10-11

100M

100K

Application Note 519

All Intersil semiconductor products are manufactured, assembled and tested under ISO9000 quality systems certification.
Intersil semiconductor products are sold by description only. Intersil Corporation reserves the right to make changes in circuit design and/or specifications at any time without notice. Accordingly, the reader is cautioned to verify that data sheets are current before placing orders. Information furnished by Intersil is believed to be accurate and
reliable. However, no responsibility is assumed by Intersil or its subsidiaries for its use; nor for any infringements of patents or other rights of third parties which may result
from its use. No license is granted by implication or otherwise under any patent or patent rights of Intersil or its subsidiaries.

For information regarding Intersil Corporation and its products, see web site www.intersil.com

Sales Office Headquarters

NORTH AMERICA
Intersil Corporation
P. O. Box 883, Mail Stop 53-204
Melbourne, FL 32902
TEL: (321) 724-7000
FAX: (321) 724-7240

EUROPE
Intersil SA
Mercure Center
100, Rue de la Fusee
1130 Brussels, Belgium
TEL: (32) 2.724.2111
FAX: (32) 2.724.22.05

ASIA
Intersil (Taiwan) Ltd.
7F-6, No. 101 Fu Hsing North Road
Taipei, Taiwan
Republic of China
TEL: (886) 2 2716 9310
FAX: (886) 2 2715 3029